signalflow-nn

Neural network extension for SignalFlow — 14 encoders, 7 heads, 4 loss functions

Part of the SignalFlow ecosystem.

PyTorch/Lightning library for financial time series classification. Provides modular encoders, classification heads, and loss functions designed for trading signal validation and prediction.

Installation

pip install signalflow-nn

Requires: Python ≥ 3.12, signalflow-trading ≥ 0.5.0, PyTorch ≥ 2.2, Lightning ≥ 2.5

Quick Start

from signalflow.nn.encoder import TransformerEncoder
from signalflow.nn.head import MLPClassifierHead
from signalflow.nn.model import TemporalClassificator
from signalflow.nn.data import SignalDataModule
import lightning as pl

# Create model
model = TemporalClassificator(
    encoder_type="encoder/transformer",
    encoder_params={"d_model": 64, "nhead": 4, "num_layers": 2},
    head_type="head/cls/mlp",
    head_params={"hidden_sizes": [32]},
    num_classes=3,  # fall, neutral, rise
)

# Create data module
dm = SignalDataModule(
    data=df,
    window_size=60,
    batch_size=32,
    split_strategy="temporal",
)

# Train
trainer = pl.Trainer(max_epochs=50, accelerator="auto")
trainer.fit(model, dm)

Encoders (14)

Classification Heads (7)

Loss Functions (4)

SignalFlow Integration

Use as a validator in the SignalFlow pipeline:

import signalflow as sf

result = (
    sf.Backtest("nn_validated")
    .data(raw=raw)
    .detector("sma_cross", fast_period=20, slow_period=50)
    .validator("nn/transformer", d_model=64, nhead=4)
    .entry(size_pct=0.1)
    .exit(tp=0.03, sl=0.015)
    .run()
)

Package Structure

License: MIT  ·  Part of SignalFlow